In: Statistics and Probability
Question 1. [7 marks total] Parental leave is often compensated to some degree, but the amount of compensation varies greatly. You read a research article that stated, "across people of all incomes, 47% of leave-takers received full pay during their leave, 16% received partial pay, and 37% received no pay." After reading this, you wonder what the distribution of parental leave payment pay is for low income households. Suppose you conduct a survey of leave-takers within households earning less than $30,000 per year. You surveyed 225 people (selected in a random sample) and found that 51 received full pay, 33 received partial pay, and 141 received no pay.
1.d) [2 marks] Compute the p-value for your test statistic and
conclude whether
you believe there is evidence against the null hypothesis in favor
of the alternative
hypothesis.
Null hypothesis :Ho :Distribution of parental leave payment pay matches with the given distribution,
Alternate hypothesis :Ha Distribution of parental leave payment pay does not match with the given distribution,
Chi-square test for Goodness of fit is appropriate to test this.
Given,
Observed counts | |
Full pay | 51 |
Partial Pay | 33 |
No pay | 141 |
Total | 225 |
Distribution of parental leave payment pay:
% leave-takers received full pay during their leave : 47%
% received partial pay : 16%
% received no pay : 37%
Expected counts :
Expected Number of people - full pay = 47% of 225 = (225x47)/100 = 105.75
Expected Number of people - Partial pay = 16 % of 225 = (225x16)/100 = 36
Expected Number people - No pay = 37% of 225 = (225x37)/100 = 83.25
Observed counts | Expected Counts | |
Full pay | 51 | 105.75 |
Partial Pay | 33 | 36 |
No pay | 141 | 83.25 |
Total | 225 | 225 |
Test Statistic :
Observed counts | Expected Counts | O-E | (O-E)2 | (O-E)2/E | |
Full pay | 51 | 105.75 | -54.75 | 2997.5625 | 28.3457 |
Partial Pay | 33 | 36 | -3 | 9 | 0.25 |
No pay | 141 | 83.25 | 57.75 | 3335.0625 | 40.0608 |
Total | 225 | 225 | 68.6566 |
Degrees of freedom : Number of categories - 1 = (3-1)=2
For 2 degrees of freedom :
As p-value : 0.0000 < :level of significance (either 0.1,0.05 or 0.001; or any) ; Reject the null hypothesis.
There is evidence against the null hypothesis in favor of alternative hypothesis.